University of Glasgow
Artificial intelligence in medicine, knowledge-based systems, knowledge modelling & refinement hypothesis.
Intelligent computer programs generally consist of a knowledge base, a reasoning mechanism, and a suitable user interface. A knowledge base is often initially constructed with the help of a domain expert and then progressively refined into a high-performance knowledge base. As knowledge can change over time, a particular interest of the artificial intelligence (AI) community are methods for automatically suggesting refinements of a knowledge base.
I am currently exploring the use of argumentation logic in the automatic refinement of computerised knowledge bases. Argumentation logic is a formal framework for modelling human collaborative deliberations, allowing the exchange of arguments in favour, or against, some conclusion based on potentially incomplete or inconsistent information.
This work is being explored in the critical care medicine domain and is particularly important as medical knowledge is rapidly changing which is challenging for many clinical decision support systems (CDSS). It is hoped that if a CDSS presents a conclusion that a clinician disagrees with, the implementation of an argumentation engine in the CDSS will allow the reasoning applied by the CDSS to be made explicit, leading to a greater understanding by the clinician, and allow the clinician to enter into a dialogue with the CDSS, suggesting ways in which both the system’s reasoning and knowledge base can be improved. This research will broadly impact on the areas of computer science, cognitive science, and critical care research.
Check out contributions by and mentions of Laura Moss on www.software.ac.uk